The hydration and isomerization of α-pinene over zeolite beta. A new coupling reaction between α-pinene and ketones

The hydration and isomerization of α-pinene over zeolite beta. A new coupling reaction between α-pinene and ketones

JOURNAL OF MOLECULAR CATA!!YSIS A CHEMICAL ELSEVIER Journal of Molecular Jan C. van der ’ DeJji lhi~r,:rity of Technology, ’ Unir,ersidade Waal A...

856KB Sizes 2 Downloads 55 Views

JOURNAL OF MOLECULAR CATA!!YSIS A CHEMICAL

ELSEVIER

Journal of Molecular

Jan C. van der ’ DeJji lhi~r,:rity of Technology, ’ Unir,ersidade

Waal

A: Chenlical

105

( 1996)

185-l 92

‘, c;fOrpnic

lahorato~

Nora de Lisboa,

Cataiysis

Chernisty

Depart~~vnmto

Received

und Cattilwis.

de QurvGx.

Jrc!imnfrrc~rr 136, 2826 BL. L?rrft. Nerl~~-fttnd.~

2KL5 Monte de Caparica.

Lisbon.

Porrugul

23 June 1995; accepted 24 August I995

Abstract The catalytic potential of zeolite H-beta in the hydrGon and isom;riration CI a-pincne was investigated. In the presence of water the main product is the monocyclic alcohol a-terpineol (48%) though selectivity towards bicyclic terpenes is higher ;han observed for sulphuric acid (26% vs. 5.5%). When the isomerization is performed in pure acetone, a new compound forms by a novel C-C coupling reaction between a-pinene and zboetone.The product i.; identified as ru-terpinyl acetone. This coupling seems to be a general reaction between o+pinene and ke!!)nes and is catalyzed exclusively by zeolite beta. Kqwards:

Zeolite

beta; o-Pincne:

Isomerization;

Hydration;

C-C

coupling

Upon subjecting a-pinene P to aqueous minerai acids, complex mixtures are obtained resulting from isomerization and hydration [ I]. The main products ( see Fig. 1) are the monocyclic terpenes a-terpineol 12, limonene nd terpinolene Minor amounts of camph (Y-and y-terpinene 9,7, cy- and @, bo_mcoi 3, y-terpineol are also formed, as -weii as 1,4- and 1,8-cineol, paracymene and tricyclene [ 231. In current industrial practice, e.g. borne01 preparation, often two- or three-step processes are applied in which a-pinene is first isomerized to camphene. Much research has been done TV develop clean processes which have high selectivity towards one of the pro starting directly from cu-pinene. 1381-l 169/96/$1.5.00 SSD/1381-1169(95)00244-S

0 1996 Elsevier

Science B.V.

All rights reserved

reaction; a-Tcrpinyl

acetone

Zeolites, microporous aluminosilicates, may have potential in improving selectivity here. The kinetic size of the a-pinene molecule as well as that of most of the bicyclic products restrict the use of zeolites to the large pore zeolites (i.e. 12membered rings, access 7-7.5 A pore diameter) like Y, beta and mordenite. Two pubiications deal with the use of zeolites in the hydration and/or isomerization of a-pinene. Nomura et al. [4] used various zeolites (Ca-Y, US-Y &a-X, ferrierite) in the hydration of very diluted aqueous solutions of a-pinene to cr-terpineol. Highest selectivity to LYterpineol was obtained using ferrierite (yield - 100°C). Selective hydration of cr-pinene to bomeol and camphene using a three phase system (water, terpene and an unspecified zeolite) has been reported by Che yields and conversions are given

&

1 4H

5

2

P

/ eb

2. I. Cutalystpreparation

6

9”

ii

JHO 14

Zeolite beta (Si/Al= 10) was synthesized according to Wadlinger and Kerr [ 61. I I .61 g sodium aluminate (0.1415 mol NaAl&, Ftiedel&I), 110 g 40% (w droxide (0.299 mol I Ludox kS i were 291.59 g silica sol ( I .461 -lined autoclave. mixed and transferred ti a After crystallization ( l/k5 was collected by filtration, dried at 115°C and calcined at 500°C to remove the organic template. -beta is obtained y ion exchange with NH!JW3 and subsequent ac ation at 500°C. Outer surface deactivation of zeolite beta by means of &a!umination, as reported earlier by Rigutto et al. [ 9 ] , was performed by overnight wa synthesized zeolite with a 0.025 solution ( 34 mlig zeolite), followed by thorough washing with distille water prior to cakinatk::. In the same way, but pplying O.Q9or 0.035 mol NaAD,, samples with Si/Al= 20 red and zeolite boron beta (Si sked according to van der Waal et al. [91. 2.2. Generul procedurejor monotcrpene

pure borneol is claimed after distillation. Mowever, in both cases relatively ong reaction times are necessary for high conversions of a-pinene (>9Oh). In this work we report on the fast hydration and/ or isomerization of cu-pinene using zeolite as the catalyst. Zeolite beta is a high-silica (WA1 10 to m) [6,9] with a three-dimensional pore system containing 12-membered ring apcrtures (7.4 X 6.5 A) [ 81, which makes this zeolitc very suitable as a regenerable catalyst in orga reactions. Upon performing zeolite beta cataly isomerization in ure acetone a new C--C cou reaction was dis vered which is described i aper and appears to be a general reaction of LYinene and ketones.

hydration or isomerization

The catalyst (0.5 g) was suspended in a solution of 1,3,5-tri-isopropylbenzene (internal standard, 1.65 mmol) in the solvent (90% aqueous acetone, acetone, butanone, cyclohexanone or dioxane) and the mixture was heated up to reaction temperature, 56°C under Iv2 atmosphere. After 30 min the reaction was started by addition of 1.65 mmol of the starting compound ( a-pinene, onene, a-terpineol or borneol; all purchased Janssen Chimica) to the reaction mixture. Samples were taken at regular intervals and ana.53 mm CP Sil-5 ntification of the

5.17. van der Waal et al. / J;lurnul

ofMolecular

Catalysis

A: Chemical

105 (1996)

185492

_--

C v

3.1. Hydration and isomcrkation

b *

of cr-pinene

a+

The hydration of a-pinene in 90% ( w/w ) aqueous acetone in the presence of -beta (Si/ Al = 10) yielded a complex mixture of monoterpenes (both alcohols and hydrocarbons). The Table

1

Sciectivity to product groups and to bomeol (3)

and a-terpineol

( It) in the isomerixation/hydration of cu-pinene A Catalyst Water Cow. added (%)

t

B-T

B-A

3

-

75

!.3 20.05

2.97

75

2.5 26.23 14.50

-


I6 2.8

7s

M-T

M-A

7.59

3.04 79.95

7.01

4.93

5.15 73.77 47.57 42.76

-

-

-

-

-

-

5.5

-

-

94.5

-

-

na!ed!; Cl-Y (FK!

mg. Si/AI=X);

HzSQ

(33 mg). in 30 ml

acetone at 56°C. alcohols;

B-T = bicyclic

terpenes;

M-

A = monocyclic alcohols: M-T = monocyclic terpenes.

25 7

v

IV Scheme 1.

s

6

main products (see Table 1) are the monocyclic terpenes, also found by catalysis with mineral acids. -beta differs from sulp ing a substantially big penes, predominantly zeolite H-beta in co

12

’ Using zeolite H-beta (200 n- !z. Si/At = IO. outer surface dealumi-

” B-A = bicyclic

E

Scheme 2.

_.

H-beta

-

.

Selectivity ( %) to h

H-beta H-Y

B

,

th)

(g)

H,SO,

187

(see Table 1), which is formed via (cf. Scheme I ) . Small amounts of borr,col3 are formed over H-beta while isoborneol2 is initially not formed. The selective fomation bomeol over isoborneol, which i_sther,nodyna icaiiy more favoured, can only tie explained direct attack of water on the pinene cation give borneol before its rearrangement to cation V which apparently rather de rotonates to give campheae than hydrates to gi isoborneol. This is in arkhasb [ 101 who states t agreement with isoborneol is formed from cation neol is formed from cation thought to rearrange fast to is not assumed to be an intermediate on which water can attack [ I 1. The protonation of (Ypinene to give cation is generally accepted to be the slowest (rate determining) step in the acid catalyzed rearrangement. 3.2. Kinetics isomerization

40

ofa-pinene hydration and

0

etic curve,t; ifo~ hydratioli. and isomeriin the presence of water can

CSi/ (w Iw) aqueous acetone. Product groups

Fig. 2. Hydration and/or isomerizationof cx-pinene over H-beta Al = IO, at 56°C) in 90% according to Scheme 3:

cr-pinene; 0 - bicychc alcohols:

bicyclic hydrocarbons; 0 - monocyclic hydrocarbons; & - :erpineels; A .-

I ,a-terpine.

steps

between each group of

Zro!i te treatintznt

A I unrxtr;lcted uwxlcined h Al

X)I

I .71

uncxtracted culcmed

B 1 extracted uncalcincd ”

OIIC

B1 extracted calcined ___--

5.97

Table 3

Zeolitc compiti~in

liatc Conam (IO

SiiAl _~.__

--

SI!H

‘g,<..!,,,,.J _--

__--.______._

5 .v7

IO ‘0 __________~__.

________..

‘dm‘b

IO 5 70

0.64 -.- ...^._._._....

- -- ..-- - ..-----

” Olltsr rurf;lcc tlc;dutninCwd bct;~, W’i ‘.I /u .Lq”‘““~ ;LI’CIlNlT Al 50°C.

Scheme 2 and Scheme 3. The difference between Scheme 2 and Scheme 3 is the number of curves used to represent different type of pseudo species, as convenient. Since under homogeneous conditions kinetics also can be described by similar models [ lOI, i.e. assuming first order elementary steps, we assume a similar mechanism for zeolite catalyzed reacrions. We note however that for a first-order reaction there is no influence of internal diffusion limitations on the apparent order of the reaction rate, so determination of the nature of the rate determining step (kinetic or difr’usion) is not possible from our experiments. For zeolite H-beta catalyzed isomerization in pure acetone the kinetics change to Langmuir-Hinshelwood kinetics though the first order in a-pinene remains.

Since the outer stir-face ofa zeolite also contains active a&J sites which do not exhibit shape-selectivity due to the absence of a pore structure, removal of such sites could enhance selectivity. This, was indeed found by Rigutto et al., in the Fischer-indole synthesis using zeolite beta as the e dealumination of the outer-euror large pore zeolites containing organic templates, can be achieved by treatment of the as-synthesized material with Na, The p,rescnce of the template protects the internal aluminium sites for removal. The effects ofoutcrsurface dealumination on the pseudo first-ok;!er rate constant of a-pinene hydration and isomerization are shown in Table 2. Previous studies on thermogravimetric analysis of tetraethylr;mmonium (TEA+ ) degradation in zeolite beta, showed that the decomposition of TEA+ starts around It is therefore assumed that in caaaiyst I (activated at 250°C, Table 2) all TEA ’ is slill present in the zeolite channels and reaction can only proceed via outer-surface aluminium sites. The high reactivity of uncalcined, ulatreated zcolite beta (catalyst A I, Table +?Isuggcsts that II large amount ofourer-surface acid sites is presoimi. The Na, TA treatment effectively rcmovl’s all outer-surface almminium as shown by eactivation of catalyst Bl. Upon dealumination and subsequent calcination, catalyst activity is greatly improved (catalyst B2) in comparison with the untreated catalyst A2. This enhanced activity may be due to removal of noncrystalline materials from the outer-surface, which blocked pore openings.

Catalyst activity changes dramatically with zcolite composition. Table 3 shows kinetic data for three different Leolites beta. Catalyst activity increases with increasing silicaE to aluminium ratio. Three cxplanations are possible: (i) the hydrophobicity ofa zeolite increases with increasing Si/Al ratio. This is due to the asid sites which

I._..

0

0.0

I

0.2

I I(I

-I-I_____

O.? 0.6 Fwrioni3l Conversion

G.8

Fig. 3. Ki’fect olzeolite composition on the sekctivity to wrerpineoi in the

hydratioo/i!;omerizationof u-pineile at 56°C in 9OV f w/w) Si/AI=10;3Si/AI=LO;OSi/B=20.

-I

OL-. 0.0

0.2

0.4

06

0.8

1 II

Fractional Conversion

Fig. 4. Effect of zeolite composition on :he selectivity to bicyclic terpenes in the hydration/isornerization

( w/w ) aqueous acetone.

SJAI=

of Lu-pineneat C6”C in 90%

IO: nSi/AI=20:

LSi/B=X

1.;

6

!d heme 3.

sion limitations. Since there is a pseudo first-order rate constant, this can, howeiier, neither be rejected nor supporred on the Si ; of our experiments. i ii) The sorption distri tion witnin a zeolite changes with increasing Si/Al ratio (due to the increasing hydrophobicity). The intrinsic concentration of terpenes in the zecJite pores wiii become higher thereby increasing reaction rate. It t-n’=: '"J be noted that (i) and (ii) ,rre in essence i) The acid strength of an alundent on the number of nextinium sites. The higher the i&boring aluminium sites the lower the acid strength. A igher Si/Al ratiotherefore means stronger aci sites together with a lower number of acid sites per gram of zeolite. Effect (iii) is perhaps the unlikeliest explanation since the difference between %/Al = 10 and 20 in absolute number of next-~eighbo~ng sites is relatively small. The large difalumi boron zeoference between the aluminium a e very low hte beta (Table 3) is explained acid strength of a proton on a boron site. There is little effect of the zeolite composition on the selectivity towards a-te ineol though the boron zeolite seems to keep a high selectivity towards almost complete conversion (see Fig. 3). Selectivity towards bicyclic alcohols is favoured by silicon to a~u~~i~iurn ratio as can be seen in 3.5. Formation of a new product, a-?erpinyl acetone

60 40 20 0 40

80

120

If@

200

m/z

Fig. 5. Mass spcstrum of a-terpinl I xe:tiiic

strongly adsorb water and polar solves’u: molecules. In case of low %/Al ratios, absorbed water and solvent molecules could beco dering the diffusion cif ar-pinene into the zeohtes pore system. This suggests one would find diffu-

Upon isomerization of a-pinene in pure acetone (a trace amount of water is assumed to be present in the zeolite or acetone used) a new product is formed. This product is not formed when dioxane is used as the solvent or when other acids (sulM-4!, amorphous phuric acid, zeolite Y, are applied as the si!ica-aluminas, and SA entification by means of mass specct to be identitied troscopy allowed the as a-terpinyl acetone This coupling was further confi tone-d, as the solvent. T deuterated a-te inyl acetone not only showed

Table 4 Effect of ketone structure on the first-or.ler rate constant of n-pinene isomerization and selectivity to a-terpinyl ketone catalyzed by Leolite H-beta * Ketone

acetone ’ acetone

-__,. Rat2 LWnsLmt0:

Sclcctivity to

a-pinene conversion

a-terpinyl ketone h

(10

[ 5% j

“dm3h‘”

g)

10.2

15.1

6.0

0.0

(90 wt% in water) ’ acetone

81.4

7.7

acetone-d,

54.6

8.9

butanone cyc1ohexanon.z

32.4 1.7

IO.0 _d

pinacolone

I.1

0.0

-beta (Sil AI = 40). 30 ml pure ketone at S6”C. ” At a-pinene conversion of > 99%. ’ Si/AI = IO used.

(

see Table 4) _ which can

hindrance of t-he bulky butyl group. The pseudo first order rate constant for a-pinene conversion is strongly dependent on the nature of ketone used d is found to decrease with increasameter of the ketones: acetone ‘> butanone > cyclohexanone > pinacolone. Kinetic modelling of the experimental data can only be achieved if an intermediate species between ar-pinene and acetone is assumed. However, such a species was not detected by GC, suggesting that this effect should be ascribed to internal diffusion li pinyl acetone. This effect is even more pronounced for the coupling of a-pinene and the other ketones.

’ Selectivity at complete Lu-pineneconversion could not be detcrmined, bumthe extrapolated yield is better than for butanone.

Schcmc 5

mole lraction Fig. 6. I:.;.~iil iate iis n function of mole fraction pinene.

evidence of a -Cl&-CO-C showed an isotro

e found that but and cyclohexanone also entered this 13ew ing reaction Qhereas pinacolone ( 3,3-di yl-Zbutanone) did not

3.6.

echmisrn

of a-terpinyl

acetorie forinution

The use of related terpenes as starting material 01) which can did not result in e. This indicates rpinyl acetclne is directly formed by a specific reaction between cr-pinene and acetone in a concertc2 way as shown in Scheme 5 rather than e cxistemcc:and stabi~i~at~~~of an intermediyate sarbocatis~ (e.g. cation ty acetone. We assume that the a-pinene molecule ;I&orbs on an acid 7ite and that the enolic form of acetone adsorbs 011 a nearby oxygen atom. T us, a protic and a neighbouring basic site are re ired. The enolic acetone could then attack the bridging carbon of the gemdimethyl group f the ar-pinene molecule. This mechanism is su orted by the kinetic behaviour of the initial rate of formation of a-terpinyl acetone, which goes through a maximum when the mole fraction of au-pinene changes from 0 to 1 $1~ fitted line represents a Langmodel, assuming a second ween acetone and a-pinene e generally product selectivity

to be sha r reactant selectivtty, ition state selectivity

thought

Si

P\

i cHxo!H)

c$ 7 Al’

0

iI

7”

0

‘Si’

A

Si

ments for the concerted mechanis in Scheme 5 are not met.

n

.A. ‘l-l

TO\ /OX 0 Al’ Si'

n

Fig. 7. Protonation of cY-pincnein mr’l~es. A - in abserxe 31.water; B - in presence of water. [ 121. Since the terpinyl ketones are only fb;rmed using zeolite beta and not with the other acid catalysts tested, the product is apparently formed by transition state selectivity [ 121: the specific adsorption and reaction of the cr-pinene-ketlone co::p!e ir! the zeolite beta channel or intersec59n. Molecular mechanics calculations on the trarjsition state show that the distance between the zeolite proton and the enolic proton should be at least 5.2 A. The average oxygen-oxygen distance in a zeolite is about 3.04-3.14 w [ 131, thus ruling out the possibility of the reaction proceeding over a single aluminium site. The fact that only zeolite beta catalyzes this new C-C coupli indicates that only zeolite beta meet requirements for this reaction, Th:: mtax! ol”;k acid Sluz anti sutYseyuentcatalysis is dependent on the presence of water. In the absence of water there is a direct interaction between the zeolite proton and C-C double bonds as reported by Kazanski [ 141 for the adsorption of ethene. In our case an cY-pinene molecule or , wi!l maybe an ion pair involving a cation be directly bonded at the aluminium site (Fig. 7A). As a consequence the kinetic behaviour is one of the Langmuir-Hinshelwood type. In the presence of water this alkene-H-zeolite interaction presumably does not exist due to preferentially adsorbed water 1151. The (Ymolecule is now protonated by the present (Fig. 7B) which spatially separates the CLIpinene and/or cations from the aluminium site and normal first-order kinetics are observed. This ht explain why in the presence of water no cyinyl ketones are formed: the spatial require-

rationlisomerizatio to monocyclic terpene and alcohols with a-terpine01 as the principal roduct (up to 48%). The selectivity towards the corn esting bicyclic products is significantly better than observed for (about 5%). Upon dealuminati outer surface with Na &DTA, an enhancement of the reaction rate a an increased selectivity towards a=terpineo! is observed. The reaction rate increasing Si/Al ratio, wbic possibly due to increased hydrophobicity of the zeolite, selectivities are however not significantly a ted. Upon isomerization of a-pinene in pure ketones (acetone, butanone or cyclohexanone) a new type of product is formed by a C-C coupling reaction of a-pinene and the ketone. Structure identification, performed by CC-MS, showed these products to be I -( a-terpinyl)-acetone, utafl-2-one aild 2-( cw-terpinyl)cyclohexanone, respectively. The a-terpinyl ketones are not formed when arher monoterpenes are used as starting reagents, suggesting that they are formed directly from a-pinene and not by any of the i~te~ediate classical or non-classical carbocations usually postulated to describe the isomerization of terpenes. A concerted mechanism is postulated in which a ketone molecule in the enolic form attacks an a-pinene molecule adsorbed on an acid site. bis is supported observation that the kin cs is first order m LYpicene as well as in ketone and by the Lang inshelwood type of kinetic bebaviour. The new reaction appears to be specific for zeoli which indicates s cific interaction a state by the pore system of zeolite beta.

[ 7 I J.C vu der Waal, MS. Rigutto and h van Bekkum. J. Chem. %c., Chem. Commun., ( 1994) 1242.

[ I ] R.M. Albert. S.G. Traynor and R.L. Webb. in D.F. Zinkel and J. Russell { Eds. ), Naval Stores. PULP Chemical Associsr~on. New York, 1989, p. 479. [ 21 G. Valkanas and N. Inconomou. Helv. Chim. Acta, 46 ( 1963) 1069. [ 31 CM Williams and D. Whittaker, J. Chem. Sot. (B ), ( 197. ) 672. M. Nomura. Y. Fujihara, H. Taka:a. T. Hirokawa and ~1. Yamada. Nippon Kogaku Kaishi, i ( 1992) 63; CA 116 12926Xm. Q. Chen, X. Cdi and S. Cao, Chin. Pat. 1.049.842,( 1991). CA 11.5:P159483z. R.L. Wadlinger and G.T. Kerr. US Pat. Appl. 28.341. ( 1975).

i 81 J.B. Higgins. R.B. LaPierre, J.L. Schenker, A.C. R0hrman.J.D. Wood, G.T. Kerr and W.J. Rohrbaugh, Zeolites, 8 ( 1988) 446. 191 MS. Rigutto. 1l.J.A. de Vries,S.P.. Magili,A.J. Hoefnagel and I-I. van Bekkum, Stud. Surf. Sci. Catal., 78 ( 1993) 417. 101 V.A. Barkhash, in Top. Curr. Chem., 1161 I17 ( 1984) 1. I I I E. Bourgeat-Lami, F. Di Renzo arid F. Fajula, J. Phys. Chem., 96 ( 1992) 31107. I21 N.Y. Chen, T.F. Degnan and CM. Smith, MolecularTransport and Reactions in Zeolites, VCH, New York, 1994, p. 173. [13 81R. de Ruiter, ThGs, Delft University of Technology, ( 1993). p. 28. 114 I V.B. Kazansky, Stud. Surf. Sci. Catal., 85 ( 1994) 260. [I5 1 A.G. Pelmenschikov, R.A. van Santen. J.H.M.C. van Wolput and J. Jlnchen. Stud. Surf. Sci. Catal., X4 ( ?994) 2 159.